BgHdr proteins were expressed as fusion proteins with thioredoxin (17.8 kDa) at N-terminal and His-tag at C-terminal. The calculated molecular mass of the fused BgHdr1 and BgHdr2 were 51.33 and 49.9 kDa, respectively, as were confirmed by SDS-PAGE (Fig. 12). Both BgHdr1 and BgHdr2 have conserved regions in their amino acid sequences. For example, Cys27, 111, and 209 in BgHdr1 and Cys12, 96, and 194 in BgHdr2 are well-conserved, because they participate in iron-sulfur cluster formation (Fig. 13). In addition, His56, 89, and 139 of BgHdr1 and His41, 74, and 124 of BgHdr2 were also highly conserved (Fig. 13) for coordination with diphosphate group (Gräwert et al., 2009). Phylogenetic analysis showed that each isozyme belongs to different bacterial Hdr clade (Fig. 14).
５３
５４ Figure 12. SDS-PAGE of the recombinant Hdr.
Lane 1, crude fused BgHdr1; Lane 2, purified fused BgHdr1; Lane 3, crude fused BgHdr2; Lane 4, purified fused BgHdr2.
Both recombinant proteins purified on Ni-NTA column.
５５
Figure 13. Homology among bacterial Hdr enzymes.
AaHDR (Aquifex aeolicus, aq1739), EcHDR (E. coli, Y75_p0029), BgHdr1 (B.
glumae, bglu_1g28650), and BgHdr2 (bglu_2g22170). The residues marked with circle (Cys27, 111, and 209 in BgHdr1 and Cys12, 96, and 194 in BgHdr2) participate in iron-sulfur cluster. In addition, His56, 89, and 139 of BgHdr1 and His41, 74, and 124 of BgHdr2, marked with triangle, coordinate with diphosphate group (Gräwert et al., 2009).
５６
Figure 14. The phylogenetic tree of bacterial Hdr.
The each BgHdr isozyme belongs to the different clade.
５７ 3. Product ratio of BgHdr
IPP and DMAPP produced from HMBPP were cleaved to give corresponding alcohols, which were subsequently separated and identified by GC. The chromatogram of the alcohols showed Rt values at 7.12 and 8.05 min, which corresponded to those of standard 3M3B1ol and 3M2B1ol, respectively (Fig. 15).
The peak area ratios of 3M3B1ol to 3M2B1ol, corresponding IPP to DMAPP ratio, were 2.28 for BgHdr1 and 2.22 for BgHdr2, indicating that the ratios were virtually identical in both isozymes (Fig. 15).
５８
Figure 15. FID-GC analysis of products of in vitro reaction of BgHdrs.
3-Methyl-3-buten-1-ol (3M3B1ol) (Rt, 7.12 min) to 3-methyl-2-buten-1-ol (3M2B1ol) (Rt, 8.05 min) ratio, corresponding to IPP to DMAPP ratio, was about 2.2:1 for both BgHdr1 (black line) and BgHdr2 (broken line).
５９ Table 5. The ratio of IPP and DMAPP.
AaHdr (Aquifex aeolicus Hdr), EcHdr (E coli Hdr), BgHdr1 (B. glumae Hdr1), BgHdr2 (B. glumae Hdr2). The product ratio of BgHdr was different from those of previous studied bacteria. Aquifex aeolicus and E. coli have only 1 hdr gene.
IPP:DMAPP reference
BgHdr1 2:1 This study
BgHdr2 2:1 This study
AaHdr 5:1 Altincicek et al., 2002
EcHdr 6:1 Gräwert et al., 2004
６０ 4. Kinetics properties of BgHdrs
To measure kinetic parameters of BgHdr, artificial electron donor MV and DT were used. MV has been used as electron donor for iron-sulfur clusters and was previously demonstrated that use of MV gave reaction kinetics of bacterial Hdr comparable to the reaction using the biological electron donor flavodoxin and NADPH (Gräwert et al., 2010, Wang et al., 2010).
Lineweaver-Burk plot of MV oxidation by BgHdr (Fig. 16) yielded kcat andKm
values of BgHdrs (Table 4). BgHdr1 had 10-fold higher kcat/Km and 3.5-times lower Km values than those of BgHdr2. This meant that two Hdr in B. glumae could differentially function under the various environmental conditions. It is possible that Hdr2 was expressed constitutively to function with high substrate concentration, while Hdr1 could be induced to cope with stress situation where carbon source is limiting.
６１
Figure 16. Lineweaver-Burk plot of BgHdr isozyme.
BgHdr1 (-●-); BgHdr2 (-○-).
The insert shows Michaelis-Menten plot.
６２ Table 6. Kinetic parameters of bacterial Hdrs
Kinetic parameters of various bacterial Hdrs. BgHdr1 was 10 folds more efficient than BgHdr2.
Km
(µM)
kcat
(min-1)
kcat/Km
(µM-1 min-1)
Specific activity
(μmol min-1mg-1) Reference
BgHdr1 6.0 ± 1.1 187 ± 6 31.1 ± 6.2 3.66 ± 0.11 this study
BgHdr2 21.2 ± 5.5 66.6 ± 5.2 3.1 ± 0.8 1.39 ± 0.11 this study
AaHdr 590 ± 60 222 ± 12 2.66 6.6 ± 0.3 Altincicek et al., 2002
EcHdr 30 - - 0.7~3.4 Gräwert et al., 2004
６３ 5. Expression pattern of hdr isogenes
To assess the temporal expression pattern of hdr isogenes during growth cycle, Northern blot analysis was performed using RNA from B. glumae BGR1 that was cultured for 10~19 h in M9 minimal medium (Fig. 17). Transcript message level of both hdr1 and hdr2 increased as culture progressed from lag (10~11 h) to exponential phase (12~14 h). After the growth reached stationary phase, the message level of hdr isogenes was dramatically reduced. The transcription was most active at the exponential phase (14 h), with higher transcription level of hdr2 than hdr1 throughout the growth (Fig. 17). These results were confirmed by transcriptome analysis of LB medium-grown bacterium (data from prof. I. Hwang, Seoul National University) (Table 5). The hdr2 gene was transcribed 1.8 times higher at exponential phase (6 h) and 1.6 times at stationary phase (10 h) than hdr1.
６４
Figure 17. The expression pattern of hdr isogenes.
a. Nothern blot results of hdr isogenes. 2μg of total RNA was loaded and blotted.
b. The RNA extraction time points after inoculation.
Both hdr genes were increasely transcribed to early stationary phase. The hdr2 was more transcribed than hdr1 in BGR1.
６５
Table 7. The transcriptome analysis of MEP pathway genes.
The RNA sequencing data of MEP pathway genes were provided by prof. I. Hwang.
The RNA of BGR1 was extracted at 6 h and 10 h after inoculation in LB medium.
Like Nothern blot analysis, hdr2 was transcribed more than hdr1.
６６
6. Growth rates of hdr isogene knock-out mutants
The hdr isogene knock-out mutants were constructed by insertion of Ω cassette using triple mating (Lee, 2006). To observe phenotype of the knock-out mutants, the growth rates were measured under the various conditions. When grown in LB medium, the performance of BGR1, HDR1KO and HDR2KO were identical (Fig.
18). However, on LB plate, the colony size of HDR1KO was smaller than the others, while the number of colonies was the similar (data not shown). The mutants were cultured under the heat shock condition at 42 ºC. HDR1KO did not grow normally and colony size difference was larger than under 37 ºC incubation (Fig. 19).
Susceptibillity of HDR1KO was again observed in culture grown at extreme pH.
The difference in the growth rate was most obvious at pH 4 and 5. At alkaline pH, all the strains failed to grow (Fig. 20). The growth of HDR2KO was not different from BGR1 against heat and pH shock. When the strains were cultured in minimal M9 medium, the retarded growth of HDR1KO was more pronounced than in rich LB medium (Fig. 21). In the case of the inoculation on rice plants, the population of HDR1KO was 10~640 fold lower than that of HDR2KO and BGR1 (Fig. 22). This lower population of HDR1KO was probably due to inability to properly colonize the plant. In swimming test, the radius of HDR1KO was smaller by 28 % than HDR2KO and BGR1. The small radius again indicated the growth defect of HDR1KO (Fig. 23).
６７
Figure 18. Growth of B. glumae starains in LB medium at 37 ºC.
Growth of both HDR1KO and HDR2KO mutants grew was not different from that of BGR1. Circle, BGR1; rectangle, HDR1KO; triangle, HDR2KO.
６８
Figure 19. Growth under the heat shock condition at 42 ºC.
BGR1 and hdr knock-out mutants were incubated at 42 ºC on LB medium.
６９
Figure 20. Growth under the pH shock condition.
Each B. glumae strains were cultured in LB medium with various pH conditions at 37 ºC. a. pH 4 buffer (100 mM MES); b. pH 5 buffer (100 mM MES); c. pH 7 buffer (100 mM MOPS); d. pH 8.5 buffer (100 mM bicine).
７０ Figure 21. Growth in M9 minimal medium.
The B. glumae strains were cultured in M9 minimal medium (0.2 % glucose) at 37 ºC. Growth of HDR1KO lagged behind those of HDR2KO and BGR1.
７１
Figure 22. B. glumae cell densities in rice plant inoculated with the bacterium.
After B. glumae was inoculated to rice seedling, the cell bacterial cell density was meassured every day for a week. The cell density of HDR1KO was 1/10 ~ 1/640 of BGR1 and HDR2KO.
７２
Figure 23. Effect of hdr knock-out on B. glumae swimming.
The swimming halo diameters were measured 36 h after inoculation on swimming plate (LB, 0.4 % agar) at 28 ºC. The reduced swimming ability of HDR1KO was caused by growth defect.
The halo radius (cm): 25.0 ± 0.8 for BGR1, 18.0 ± 0.9 for HDR1KO, and 25.0 ± 0.0 for HDR2KO.
７３ 7. Virulence of the knock-out mutants
The pathogenicity of the knock-out mutants was observed on rice plants at seedling and flowering stages. Rice seedling inoculated with BGR1 and HDR2KO developed wilt symptom accompanied by leaf browning. However, HDR1KO did not exhibit such symptom (Fig. 24). In the case of inoculated rice panicles at flowering stage, HDR1KO-inoculated rice exibited low of grain rot symptom (disease index 1.5), compared to the high pathogenicities of BGR1 (disease index 3.2) and HDR2KO (disease index 3.5) (Fig. 25). The low virulence of HDR1KO was compatible with the previously described low bacterial cell density (Fig. 22) of the strain upon inoculation.
７４
Figure 24. Rice seedlings inoculated with hdr knock-out mutants.
７５
Figure 25. Rice panicle inoculated with B. glumae hdr mutants at the flowering stage.
７６
8. Identification of cellular proteins modified in hdr isogene knock-out mutants
After the hdr isogene knock-out mutants and BGR1 were cultured in M9 medium, Cellular proteins were extracted from the strains at 6 h (exponential phase) and 10 h (stationary phase) after inoculation on M9 medium, and eletrophoresed on SDS PAGE. It was outstanding that 3 protein bands with apparent molecular mass of 36, 52, and 57 kDa of HDR1KO migrated faster than the corresponding bands of HDR2KO and BGR1 (Fig. 26). The bands were isolated, digested by trypsin, and identified by LC MS/MS analysis (Table 8). It was shown that the identities of proteins were the same in the corresponding bands. The most abundant protein, with molecular mass of 57 kDa, was identified as GroEL with 90 % coverage of peptide against Database (bglu_1g07150, bglu_2g19330) (Fig. 26). The protein band with 57 kDa reacted with GroEL antibody in Western blot (Fig. 27), further confirming the identity of the protein. The LC MS/MS analysis of GroEL band in HDR1KO harvested at exponential phase suggested that m/z value of Lys390 increased by 42, suggesting acetylation at the residue. .In contrast, GroEL of BGR1 and HDR2KO were not acetylated at this position (Fig. 28). However, at stationary phase, acetylation at Lys390 residue had also occured in BGR1 and HDR2KO.
７７
７８
Figure 26. SDS-PAGE of cellular proteins at exponential (6 h) and stationary (10 h) phases.
The cellular proteins were separated on 10 % acrylamide gel. The yellow numbers indicate the analyzed protein bands by LC MS/MS.
a. SDS PAGE of cellular proteins at 6 h after inoculation. The 3 bands of HDR1KO migrated faster than those of other strains. The orange arrows indicate the isolated bands to LC MS/MS analysis.
b. SDS PAGE of cellular proteins at 10 h after inoculation. Expression of some proteins changed. The migration of 4 bands changed in comparision with protein migration pattern at 6 h. The 57 kDa protein of all strains migrated equally. The orange arrows indicate the isolated bands.
c. Among different migration bands, the 57 kDa protein was analyzed as GroEL.
The coverage numbers means peptides coverages in LC MS/MS analysis. The GroEL of HDR1KO migrated faster than that of strains at exponential phase. But the difference of migration speed was disappeared at stationary phase, due to the faster migration of BGR1 and HDR2KO.
７９
Table 8. Identification of isolated cellular proteins from bands
The cellular proteins were isolated from bands (B, BGR1; 1, HDR1KO; 2, HDR2KO) (Fig. 26). The proteins were analyzed by LC MS/MS.
80
Band name Exponential phase (6h) Stationary phase (10h)
B-57 kDa GroEL (bglu_1g07150, bglu_2g19330) GroEL (bglu_1g07150, bglu_2g19330) 1-57 kDa GroEL (bglu_1g07150, bglu_2g19330) GroEL (bglu_1g07150, bglu_2g19330) 2-57 kDa GroEL (bglu_1g07150, bglu_2g19330) GroEL (bglu_1g07150, bglu_2g19330)
B-52 kDa
ATP synthase F1, beta subunit (bglu_1g00770) ATP synthase subunit A (bglu_1g00750) S-adenosyl-L-homocysteine hydrolase (bglu_1g01990)
Isocitrate lyase (bglu_1g12420)
S-adenosyl-L-homocysteine hydrolase (bglu_1g01990)
1-52 kDa
ATP synthase F1, beta subunit (bglu_1g00770) ATP synthase subunit A (bglu_1g00750) S-adenosyl-L-homocysteine hydrolase (bglu_1g01990)
Isocitrate lyase (bglu_1g12420) Histidine ammonia-lyase (bglu_1g25210)
2-52 kDa
ATP synthase F1, beta subunit (bglu_1g00770) ATP synthase subunit A (bglu_1g00750) S-adenosyl-L-homocysteine hydrolase (bglu_1g01990)
Isocitrate lyase (bglu_1g12420)
S-adenosyl-L-homocysteine hydrolase (bglu_1g01990)
B-36 kDa ABC-type phosphate transport ABC-type phosphate transport
81
system periplasmic component (bglu_1g11450) system periplasmic component (bglu_1g11450) Translation elongation factor Ts (bglu_1g12730)
1-36 kDa ABC-type phosphate transport
system periplasmic component (bglu_1g11450)
ABC-type phosphate transport system periplasmic component (bglu_1g11450)
2-36 kDa ABC-type phosphate transport
system periplasmic component (bglu_1g11450)
ABC-type phosphate transport system periplasmic component (bglu_1g11450) Translation elongation factor Ts (bglu_1g12730)
B-44 kDa ABC-type sugar transport system periplasmic component
(bglu_1g08070)
1-44 kDa ABC-type sugar transport system periplasmic component
(bglu_1g08070)
2-44 kDa ABC-type sugar transport system periplasmic component
(bglu_1g08070)
82
Figure 27. Western blot of cellular proteins with GroEL antibody.
The proteins, extracted from B. glumae at exponential phase, were separated by SDS-PAGE and was allowed to react with GroEL antibody.
A. E. coli GroEL .
B. B. glumae protein band near 57 kDa.
83
84
Figure 28. Mass spectrum of 57 kDa-protein isolated from HDR1KO at exponential phase.
The decapetide fragment containing Lys390 (circled) is shown to be modified by acetylation as suggested by increase of m/z by 42 at Lys390.
a. HDR1KO GroEL. b. BGR1 GroEL.
85 9. Promoter switch experiment
The vector pLAFR6 was engineered to harbor 6 promoter::ORF combinations between 3 putative hdr promoters (1P: 300 bp upstream of hdr1 translation starting point, 1Po: 576 bp upstream of slp, 2P: 327 bp upstream of hdr2 translation starting point) and 2 hdr ORFs (1O and 2O for hdr1 and hdr2, respectively) (Fig. 29). In practice, the putative operon promoter 1Po was used in tandem with slp to ensure polycistronic transcription of the operon. The vectors were transformed into HDR1KO, and then cellular proteins were analyzed by SDS PAGE. It became evident that the vector including putative hdr1 operon promoter (1Po), regardless of identity of the hdr isogene, could only rescued the fast moving GroEL to wild type protein (Fig. 30). In other words, acetylated GroEL returned to deacetylated form upon complementation. The complementation experiment was further confirmed by hypersensitivity reaction (HR) test on tobacco leaf. HDR1KO that did not show HR became HR-active upon complementation with vector containing 1Po. However, hdr2 promoter (2P) exerted little effect on HR (Fig. 31).
86
87
Figure 29. The construction of 6 vectors each harboring combination of hdr promoter and ORF.
a. Two putative promoters for hdr1.
b. Putative hdr2 operon
c. Combination of promoters and hdrs cloned into pLAFR3. 1P1O, putative hdr1 promoter and hdr1 ORF; 2P2O, putative hdr2 promoter and hdr2 ORF; 1P2O, putative hdr1 promoter and hdr2 ORF; 2P1O, putative hdr2 promoter and hdr1 ORF; 1Po1O, putative hdr1 operon promoter and hdr1 operon; 1Po2O, putative hdr1 operon promoter and hdr2 operon.
88
Figure 30. Electrophoretic mobility change of GroEL cuased by complementation of HDR1KO with 6 combinatorial vectors.
a. Complementation of DRR1KO with the promoter-hdr combination could not rescue the shifted GroEL mobility.
b. Combination of 1Po with either hdr could complement the HDR1KO in GroEL mobility.
89
90
Figure 31. Hypersesitivity reaction test after complementation of HDR1KO with 6 combinatorial vectors.
Complementation of HDR1KO with the hdr genes under the influence of 1Po could rescue the HR
a. HDR1KO; b. water; c. BGR1; d. HDR2KO; e. 1P1O; f. 1P2O; g. 2P1O; h. 2P2O;
i. 1Po1O; j. 1Po2O
91
10. The seperated hdr isogene regulator and signal cascades.
To identify separate signal cascade respectively leading to transcription of hdr1 and hdr2 operon, the sequence of the putative promoter region of hdr isogenes were analyzed. In the putative promoter sequence of hdr1, there was no previously reported transcriptional factor-binding motif. However, in the putative hdr2 promoter, 5 repeating ToxR binding sequences (T-N11-A) were identified (Fig. 32).
To determine whether ToxR binds to 131 bp-long putative hdr2 promoter region, yeast-1-hybrid assay (Y1H) and Electrophoretic mobility shift assay (EMSA) were performed. The yeast harboring the vectors pHIS2HDR2-131 and pGADT7ToxR could survive, regardless concentration level of AT (3-amino-1,2,4-triazole, his gene inhibitor) (Fig. 33). Furthermore, recombinant ToxR was shown to bind to hdr2 promoter in EMSA (Fig. 34). It was shown that the hdr2 gene was down-regulated in the ToxR knock-out mutant than BGR1 by RT-qPCR (Fig. 35). Therefore, it was suggested that transcription of hdr2 was regulated by ToxR independent of hdr1.
92
Figure 32. ToxR binding sites in putative hdr2 operon promoter region.
The LysR-type regulator ToxR recognizes the specific T-N11-A motif (underlined) and binds with that motif (Kim et al., 2009). The motif was repeated 5 times in the hdr2 promoter. Nucleotides of GGAG in purple circle indicates is the ribosome binding site (SD = Shine-Dalgarno sequence). The promoter of hdr1 did not have ToxR binding motif.
93 Figure 33. Yeast 1 hybrid assay.
a. The 329 bp-long putative hdr2 promoter region was cloned into pHIS2 vector to construct pHISHDR2-131.
b. The ToxR ORF was cloned into pGAD7-Rec2 (pGADToxR).
c. The HIS3 inhibitor AT (3-amino-1,2,4-triazole) was added at 0, 20 and 40 mM.
The yeast harboring both pHISHDR2-131 and pGADToxR could survive regardless of AT concentrations. The yeast only with pHISHDR2-131 or pGAD7-Rec2 could not survive without histidine.
94 Figure 34. Electrophoretic mobility shift assay.
The recombinant ToxR protein was overexpressed by fusion protein with His6 tag.
The 131 bp putative hdr2 promoter DNA was amplified by PCR and radio-labeled.
The mixture of DNA and ToxR was separated on 5 % polyacrylamide gel by the electrophoresis and recorded on BAS-2010 ( Fujifilm).
95
Figure 35. Transcript level of hdr2 in BGR1 and ToxR knock-out mutant (TOXRKO).
The transcription of hdr2 gene in ToxR knock-out mutant was down-regulated compared to that of BGR1 as shown by RT-qPCR analysis.
96
11. RNA sequencing analysis of hdr isogene mutants
To see consequence of hdr knock-out on a global scale, change in transcript distribution was assessed by Illumina sequencing. At first, genes associated with terpene biosynthesis were analyzed (Table 9). The significantly up-regulated gene was cmk in HDR1KO and down-regulated was squalene cyclase 2 (sqc2). However, in the HDR2KO, up-regulated genes were cmk, hds, undecaprenyl diphosphate synthase (ups) and undecaprenyl diphosphate phosphatase (upp), while dxs2, squalene synthase (sqs) 1 and 2, and sqc2 were down-regulated. The most up-regulated gene in HDR1KO was toxB, toxA and toxC that constitute the toxoflavin biosynthesis operon. In the case of HDR2KO, S-adenosylmethionine decarboxylase proenzyme (SAM-dcase), stress response involved genes (superoxide dismutase, chaperonin Cpn10, cold-shock DNA-binding domain protein, co-chaperonin GroES, Heat shock protein and HSP20 family protein) and ribosomal genes (30S ribosomal protein S13, 50S ribosomal protein L36, L10, L32 and L34) were up-regulated (Table 10). The data suggested that toxoflavin biosynthesis-related genes were up-regulated in HDR1KO whereas stress associated and ribosomal proteins were upregulated transcribed in HDR2KO.
97
Table 9. Transcriptome analysis of the genes involved in terpene biosynthesis.
The color indicates the changes in expression level (red, 50~100 % increase; pale red, 25~50 % increase; white, -25~25 % increase; pale blue, 25~50 % decrease;
blue, 50~100 % decrease). The RNA was cultured in M9 medium and extracted at exponential phase.
98
Table 10 The most up-regulated genes in hdr isogene mutants.
RNA was extracted at exponential phase.
99 12. Extracellular toxoflavin producticity
To confirm enhanced toxoflavin biosynthesis, toxoflavin production was semi-qunatitatively measured in hdr mutants and BGR1 by thin-layer chromatography (Fig. 36). The HDR1KO produced increased amounts of toxoflavin compared to the wild-type and HDR2KO. Fervenulin production was also increased in HDR1KO but not reumycin. The concentration of intracellular toxin was not changed (Fig. 36).
100
Figure 36. The production of toxoflavin in BGR1 and hdr isogene mutants.
The toxoflavin was extracted and separated after 24 h culture in LB medium.
a. TLC of extracted toxins The spots were visualized under the irradiation at 365 nm.
F, Fervenulin; T, Toxoflavin; R, Reumycin.
b. The toxoflavin contents measured at 260 nm.
101
DISCUSSION
Cartain bacterial genus, comprising free-living to host-specific species, has 2 hdr genes (Fig. 14). Despite genome size reduction occurred during evolution towards host-specific bacteria (Song et al., 2010), 2 hdr genes have been retained in the chromosome of B. glumae. Phylogenetic tree of bacterial two-gene Hdrs shows that there are three major Hdr clusters. For Burkholderia, two Hdrs are distributed in two different clades (Fig. 14). Interestingly, the interspecies homologies of Hdr sometimes are lower than that of intraspecies homologies.
Excluding 15 extra amino acid residues at the N-terminal position of BgHdr1, two BgHdrs bear only 56.1 % identities (Fig. 13). The phylogentic data and this low homology might imply that one of two hdr genes has been acquired by ancestral Burkholderia through horizontal gene transfer (Dutta et al., 2002).
Then why two Hdrs were retained during speciation of the genus Burkholderia. Do they have the same catalytic function? To understand the function of each Hdr, ORF of each hdr gene was isolated. The isolated genes complemented LytB (or Hdr) null E. coli mutant Therefore, both BgHdrs had identical catalytic function (Fig. 10, 11).
In vitro enzyme assay also indicated that both Hdrs are catalytically active (Fig. 15).
In vitro enzyme assay also indicated that both Hdrs are catalytically active (Fig. 15).